Wish Design Options worked

[petervanko]

....hmmmm...coming out of DD and I have a massive set of VE options I would love to try. That is if I could.

This is a complex building in that it is primarily brick (stacked wall with lots of profile editing and layer extension) and many angles. As well, I have a hilly site and changes in grade have forced me to step foundations often. Structure is modeled completely. Get the idea? I have a very complete model.

Design Options are the logical choice to explore variuous VE exercises. However, I cannot get building components from one design option into another--I keep getting errors like "cannot extend vertically complex wall." It lets me add the same model work into some options but not others! I get millions of "An insert in the main model cannot be hosted by an element in a Design Option" despite it not hosting anything--the wall has no holes, sweeps, etc.!!!

Is there a point where a building has progressed too far to use design options??? Where it just makes more sense to have separate files?

[twiceroadsfool]

....hmmmm...coming out of DD and I have a massive set of VE options I would love to try. That is if I could.

This is a complex building in that it is primarily brick (stacked wall with lots of profile editing and layer extension) and many angles. As well, I have a hilly site and changes in grade have forced me to step foundations often. Structure is modeled completely. Get the idea? I have a very complete model.

Design Options are the logical choice to explore variuous VE exercises. However, I cannot get building components from one design option into another--I keep getting errors like "cannot extend vertically complex wall." It lets me add the same model work into some options but not others! I get millions of "An insert in the main model cannot be hosted by an element in a Design Option" despite it not hosting anything--the wall has no holes, sweeps, etc.!!!

Is there a point where a building has progressed too far to use design options??? Where it just makes more sense to have separate files?

Ive found they work great even post-CD phase, when our Client is considering design changes after the fact, but before construction (and sometimes during, LOL). Getting all of the pieces you would like inside the various options is sometimes a daunting task, especially with a lot of hosted elements... But once i manage to get all of the pieces in, they work great. :)

Ive noticed i get that msg more with hosted items that penetrate the wall, and not at all with hosted elements that are simply attached to the wall (light fixtures, etc). For me, wall sweeps are a constant source of frustration as far as DO go, because one continuous sweep soemtimes wraps an entire building, and thus, its hosts are all of the walls.

But to answet your question, i dont believe a Revit mode is ever "too far" to make use of them. Our client is always designing (moreson in CD phase than earlier) and we have been using Design Options religiously (Workset enabled multi phase mall, around 220 MB compacted).

[petervanko]

Thanks for the confirmation of the difficulty if nothing else. So, what is the solution? I have lots of sweeps and things like you mentioned....

I'm tempted to copy the entire building into each design option. Is this crazy?

[twiceroadsfool]

As i am not familiar with your project, i cant say what is and isnt crazy, lol. Our project was one existing building, with extensive demolition to create a Court that meanders between 7 smaller buildings. I know in bizarre cases i have pretty much encompassed one of the smaller buildings in to a design option. BUT, for the most part i just diligently hunt and peck for the items and related hosts that are necessary to complete the design option.

FOR ME, this gets most frustrating when items are attached to roofs, lol...

I did mention in a Product Comment (http://usa.autodesk.com/adsk/servlet/index?siteID=123112&id=1109794) that i would like Revit to auto-select all related and/or necessary objects if i choose to pick an item for a DO which either hosts or has hosts that must be included. A warning or prompt may be necessary, but i would prefer that to it making me guess as to the pieces.

Aaron Rumple (i believe) pointed out to me that Revit sometimes makes these decisions as necessary, like with the hosted elements which pierce the host, and the ones that dont. Its pretty intuitive, i would just like it to go one step further, and provide me with the option to include whatever elements are holding me up.

[truevis]

...I'm tempted to copy the entire building into each design option. Is this crazy?I'd try it as long as it doesn't bring the computer to its knees. You will, presumably, be getting rid of the unwanted options at some point, anyway.

[hand471037]

I once asked one of the senior programmers at Revit about this issue, and they said that you can more or less have as many design options as you want within a project. They don't slow things down very much, for it's my understanding that unless they are visable Revit isn't thinking about them when it's redrawing a view or while you're working on something that isn't related to those options.

I've used Design Options heavily for years without trouble. However, when you've got multipule things connected within or even between Options is when things get tough and problems arise. My general rules of thumb for this are:

1. Anything that I need to put into an Option that's hosted also gets it's host (or part of that host) added to that Option. So, for example, if I need to swap between two windows that are in two Options, I'll cut out that little bit of wall that hosts them with the Split tool and put that into each Option as well instead of the whole wall.

2. Try to add 'stubs' to where the Option hits the rest of the model. For example, if I was modeling two different entryway Options, I'd again split the wall about 6" away from the rest of the model to create a 'stub' for it to connect to, and then put the rest of the whole entry into different Options. This makes for more control over the Option, and keeps problems from coming up when elements by the Option move around.

3. Start with a viable model. For example, I might model one version of a roof. Then I'll need to do some Options of it, so I'll first 'stub' the connection between the roof and the rest of the walls below by splitting the walls horizontally, and then copy the roof and stub walls into several different Options. Then I would edit the Options to look the way I want, even if that's really, really different from what I started with. This is in contrast to simply modeling from scratch within each Option.

4. Remember that you can and actually need to Copyclip, Cut, and Paste elements from Option to Option.

5. Create overall working views for each Option. Me, I make these Axon views, overall 3D views, that allow me to double-check what's in which Option and to fix problems.

With this said wall sweeps don't work so hot in options. It gets complex. If I needed to swap between several of them, I'd consider making them non-hosted (i.e. an In-Place Family) instead. Then they would swap no problem. Are you still moving walls around enough to make the Hosted ones worth the trouble?

[iru69]

Design Options do work amazingly well, it's just that many users find them non-intuitive, difficult to use, and generally more trouble than they're worth.

I used Design Options quite a bit at first, but ultimately I found they were just too much of a drag. Now I almost always just make a copy of the file and do alternatives that way.

The way I imagine Design Options being redesigned is kind of like a version control system where you could save "version" points within a single file and then create views based on that version. You could copy and paste elements from one version to another version. In the View Properties dialog box, there would be a field with a pull down where you could select what version point you want to display. You could merge versions together. Making version changes to the model wouldn't be any different than working on a model is now.

[hand471037]

The way I imagine Design Options being redesigned is kind of like a version control system where you could save "version" points within a single file and then create views based on that version. You could copy and paste elements from one version to another version. In the View Properties dialog box, there would be a field with a pull down where you could select what version point you want to display. You could merge versions together. Making version changes to the model wouldn't be any different than working on a model is now.

I totally agree with you here! I wish that the Workset stuff was extended into a full source code control system, ala Subversion, with trees, check-outs, and all, so that you could version everything and swap elements and oh man... that would rock! I have no idea how they (Autodesk) would pull it off, but it would be keen.

[petervanko]

Don't get me wrong...DOs have served me well for a long time, but this instance has brought my productivity to a standstill. The "tree" idea is fantastic, but I'm not sure it solves the dependency issues; please elaborate. I think Twiceroadsfool nailed it: "that Revit should auto-select all related and/or necessary objects for a DO which either hosts or has hosts that must be included." Because--as of now--I have no ability to test options on the model because I am unable to select the right objects (and the lack of consistency is maddening). Will let you all know how this goes...I am going to put the entire building in each design option....

...will try the "stub" idea sometime, too...

BTW: I once had 27 option sets (w/ at least two options in each) for a 37000 SF building, and experienced no slowdown. Granted, they were small moves, a furniture option here, a window here or there...so on, so forth

[hand471037]

Peter, when I'm saying 'tree' I'm using terms that come from source code control tools, such as CVS or Subversion. These are tools software developers have had for decades that allow for teams to work together and for the project to version everything efficiently. Think Worksets, but much more robust and complex. From talking to someone who worked for Revit back in the day, Worksets in Revit are directly inspired by one such tool.

This allows for real change tracking, trying out different versions of things without having to erase anything, rolling new code into old, old into new, and all sorts of really great (if not a little mind-bending) techniques for managing a huge, complex project with lots of parts that all work together. Kind of sounds like our trade huh?

However, with software code it's a simpler problem: it's just text. It's really easy for a computer to look at plain text and see what's changed, to track all the changes someone made, and to swap in and out different versions of that text into the project. Then it's not like that stuff is 'live' at all, until you compile and/or run the program, that text is static. With Revit, you're talking about live object-orientated binary models instead, and I think that the whole problem becomes much more complex...

[Andre Baros]

"I'm going to put this whole building into a design option"
That's almost as far as you have to go some times, but it works. I've found that the easiest way to deal with design options later in the game is to just add more to the options. I've also gotten into the habit of create an "empty" design option as the default so that nothing in the main model gets locked to objects in the design options.

Most of our design options happen after a lot of work has already gone into the model and they're the only way to keep the team working and still keep up with clients.

Trees for options and better selection tools would definitely help!!!

[petervanko]

Thanks for the confirmation!

This seems to be the only way to manage things once the building has gone this far. I have so many dependencies built-in: structural beam systems that frame sloping roofs use the underside of the roof as their work plane; roofs reference walls for their footprint; walls reference major control ref planes that have been pinned; tons of custom profile work on sloping walls; sweeps--which align so nicely around the entire building--unfortunately tie the entire thing together. In some ways, the parametric side of Revit is working against us here...

[petervanko]

FAiled!@!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!

Revit won't do it. It says there is a circular chain of references, and after searching "By element ID" the objetcs that are hanging it up are evidently invalid. I can't move...I'm wasting time with the tool rather than doing work...exporting to ACAD for these studies.

Thanks for all of the great posts, advice, and time to all of you who responded.